AI Makes Average Design Cheap — That's Why Taste and Strategy Are More Valuable Than Ever

The cost of producing a decent-looking interface just collapsed. And that is the best thing that could have happened to designers who actually think for a living.

AI can now generate layouts, suggest color palettes, write microcopy, and produce component variations faster than any human. The floor for purely visual quality has risen dramatically. A passable visual design is now nearly free to produce. But here is the part most conversations about AI and design skip over: passable was never the goal. Passable is what your users tolerate while they search for something better.

The real question is not whether AI will replace designers. It is whether the executives making staffing decisions understand what designers were supposed to be doing in the first place.

What Does the Data Actually Tell Us?

The numbers paint a more nuanced picture than either the doomsayers or the cheerleaders want to admit. Figma's AI Report found that 59% of developers use AI tools for core coding work, while only 31% of designers use AI for core design tasks. That is not because designers are technophobes. It is because the hard parts of design are not the parts AI is good at.

Even more revealing: only 54% of designers report that AI improves the quality of their work. Compare that to the development side, where AI tools like Copilot and Cursor have become genuine accelerators for writing and reviewing code. The asymmetry tells us something important. Code has clear success criteria (it compiles, it passes tests, it performs). Design success is contextual, relational, and deeply human. A layout that looks correct can still feel wrong. A flow that tests an isolated screen can still confuse people in the real world.

Meanwhile, "design skills" and "design thinking" have become the number one most in-demand skill category in AI-related job postings, ahead of coding. Read that again. The companies building AI products need design thinking more than they need more engineers. The demand is not for people who can push pixels. It is for people who can make decisions about what should exist, for whom, and why.

Why Is AI-Generated Design "the New Stock Photography"?

One Reddit commenter described AI-generated interfaces as "the new stock photography," and that framing is hard to improve on. Stock photography is not ugly. It is competent, available, and generic. It fills a space without filling a need. You recognize it instantly, not because it is memorable, but because it looks like everything else.

This is precisely what happens when AI generates design at scale. The output converges on the mean. It reflects the patterns in its training data, which means it produces what has already been produced. The result is a market flooded with interfaces that look fine, feel familiar, and inspire nothing. Functional but forgettable.

For certain applications, that is perfectly adequate. Internal tools, MVPs, quick prototypes: there are legitimate contexts where "good enough" is good enough. We are not arguing that every interface needs to be a masterpiece.

But for products competing for user attention, loyalty, and willingness to pay, convergence is a strategic liability. When every competitor can generate the same polished surface in an afternoon, the surface stops being a differentiator. What differentiates is the thinking underneath.

What Can AI Not Replicate?

AI is exceptional at pattern matching. It is not capable of pattern breaking. And the work that separates products people tolerate from products people love almost always involves knowing when to break the pattern.

Consider what goes into a genuinely great product experience. It requires understanding the messy, contradictory reality of human behavior: what people say they want versus what they actually do, what they need in the moment versus what serves them over time. It requires making tradeoffs between competing priorities (speed versus thoroughness, simplicity versus capability, delight versus restraint) with full awareness of the business context and the user's emotional state. It requires taste.

Taste is not aesthetic preference. It is the accumulated judgment that comes from years of research, iteration, failure, and close observation of real people using real products. It is knowing that a healthcare application needs to feel calm and certain even when the underlying process is complex. It is recognizing that a booking flow loses half its users not because of a visual problem but because of a trust problem. It is understanding that removing a feature can be a bigger design decision than adding one.

We saw this firsthand when we helped Choice Hotels redesign their booking experience. The 50% reduction in booking churn did not come from making the interface prettier. It came from research that revealed where users lost confidence, where the flow introduced unnecessary friction, and where the information architecture was working against people's mental models. An AI tool could have generated a dozen attractive booking page variations. None of them would have identified the actual problem.

Does This Mean AI Is Useless for Design Teams?

Not at all. The designers and teams getting real value from AI are using it for the parts of the process that were already low-leverage: generating initial explorations, resizing assets, producing copy variations for testing, automating handoff documentation. These are genuine time savings that free designers to spend more hours on research, synthesis, and strategic decision-making.

The mistake is confusing the accelerator for the engine. AI can compress the production phase of design. It cannot compress the understanding phase. It cannot sit in a room with a frustrated clinician and grasp why a 200-screen tablet application needs to feel effortless despite its complexity. It cannot watch a first-time user abandon a form and intuit whether the problem is cognitive load, unclear language, or misplaced anxiety. That work requires human presence, human empathy, and human judgment.

When we worked with Matrix Medical to redesign their in-home health assessment app, the path to $10 million in annual labor savings ran through deep collaboration with clinicians, not through interface generation. The design decisions that mattered most were invisible on screen: how to sequence assessments so they matched the natural flow of a home visit, how to reduce documentation burden without sacrificing data quality, how to make a complex regulatory environment feel manageable to someone holding a tablet in a patient's living room.

What Should This Mean for How You Build Product Teams?

If you are an executive making decisions about design investment, the AI revolution should be clarifying, not confusing. Here is what it clarifies.

The value of a design team was never primarily in production speed. It was in the quality of decisions they make before and during production. Research, strategy, information architecture, systems thinking, and experience design are the capabilities that compound over time. They are also the capabilities that AI cannot substitute.

This means the right response to AI is not to shrink your design team and hand Figma's AI features to a developer. The right response is to invest more deliberately in the strategic end of the design spectrum: more user research, more service design, more experience strategy, more collaboration between design and business leadership.

Organizations that treat AI as a reason to commoditize design will get commoditized results. They will ship products that look like everything else, feel like everything else, and perform like everything else. Organizations that treat AI as a tool that frees their best designers to focus on harder, higher-value problems will build products with genuine differentiation.

We have seen this pattern before, long before AI entered the conversation. When we helped Acclaris simplify complex HSA experiences for 1.4 million account holders, the 10x increase in enterprise sales revenue did not come from better-looking screens. It came from a fundamentally better understanding of how real people think about health savings accounts, translated into an experience that made the complex feel simple. That kind of strategic clarity is not something you automate. It is something you invest in.

Where Does Taste Come From?

Taste is not a mystical quality that some designers have and others do not. It is the product of rigorous practice: conducting research, testing assumptions, studying what works and why, building empathy for specific users in specific contexts, and making thousands of informed decisions over years of practice. It is, in other words, exactly the kind of skill that becomes more valuable as the cost of execution drops.

Think of it this way. When desktop publishing made it cheap for anyone to produce a newsletter, the value of great editorial judgment went up, not down. When stock photography made it trivial to illustrate a blog post, the value of original visual storytelling went up, not down. When AI makes it cheap to generate a competent interface, the value of knowing what the right interface should be goes up.

The split is already happening. On one side: faster, cheaper, more generic. On the other: more strategic, more researched, more differentiated. Both sides will find buyers. But only one side builds products that people remember, return to, and recommend.

So What Do You Do Now?

If you are evaluating how AI should change your design operations, start by asking what your design team spends its time on today. If most of their hours go to production (pushing pixels, creating variations, building specs) then AI will genuinely transform their workflow, and you should embrace that. Free those hours.

But then ask the harder question: where do those freed hours go? If they go back to the business as cost savings, you have optimized for efficiency. If they go into deeper research, better strategy, and more rigorous experience design, you have optimized for differentiation.

When user success is at the forefront of those decisions, business success follows. The companies that will win the next decade of product design are not the ones that produce the most interfaces the fastest. They are the ones that understand their users deeply enough to know what to build, and have the taste to know how to build it well.

AI made average design cheap. That is a gift, if you know what to do with the space it creates.